Toxicity accounts for approximately 20% of drug attrition, of which nearly one third is attributed to cardiovascular issues, in particular arrhythmias.1 The cost of bringing a new drug to market can exceed $1.2 billion and require more than 10 years of research. Thus, it is critical to identify cardiotoxicity early in developmen. Current in vitro screening assays for pro-?arrhythmic drug effects focus on measuring inhibition of the hERG potassium channel, which has been linked to potentially lethal Torsades de Pointes arrhythmias (TdP).2 The hERG assay, however, lacks high sensitivity and specificity: not all QT prolongation is due to block of hERG alone and not all hERG blockers result in QT prolongation or induce TdP. To address these limitations, regulatory bodies have created the Comprehensive in vitro Proarrhythmia Assay (CiPA) initiative2, which proposes use of cardiomyocytes derived from human stem cells as an in vitro model for arrhythmogenic drug potential. Establishment of this model for toxicity screening requires detailed characterization of the cardiomyocyte action potential (AP) and calcium transient (CT), the response of APs/CTs to drugs, and the correlation to clinical outcomes in humans. High-?throughput tools needed for performing these measurements, however, have been lacking. Here we propose to develop a high-?throughput, all-?optical electrophysiology platform for cardiotoxicity screening in sem cell-?derived cardiomyocytes. With our Phase I award, we demonstrated that the Optopatch platform could be used to detect changes in the electrophysiological characteristics of human derived CMs following both acute and chronic drug treatment, albeit with the throughput of a single well per recording. To highly parallelize these measurements, we propose building a 96-?well plate Optopatch instrument for simultaneous recording of voltage and calcium waveforms under paced conditions from 24 wells. This geometry will provide nearly two orders of magnitude improvement in throughput of our assay. Optopatch constructs will be optimized to allow for incorporation of the actuator and reporter proteins in each cell. We will utilize this platform to screen 50 compounds with known risk scores for Torsades de Pointes in different sources of human cardiomyocytes and use this data to develop a predictive algorithm of arrhythmogenicity.